Multidimensional Correlation Research Articles

Correlating sparse sensing for large-scale traffic speed estimation: A Laplacian-enhanced low-rank tensor kriging approach

Traffic speed is central to characterizing the fluidity of the road network. Many transportation applications rely on it, such as real-time navigation, dynamic route planning, and congestion management. Rapid advances in sensing and communication techniques make traffic speed detection easier than ever. However, due to sparse deployment of static sensors or low penetration of mobile sensors, speeds detected are incomplete and far from network-wide use. In addition, sensors are prone to error or missing data due to various kinds of reasons, speeds from these sensors can become highly noisy. These drawbacks call for effective techniques to recover credible estimates from the incomplete data. In this work, we first identify the issue as a spatiotemporal kriging problem and propose a Laplacian enhanced low-rank tensor completion (LETC) framework featuring both low-rankness and multi-dimensional correlations for large-scale traffic speed kriging under limited observations. To be specific, three types of speed correlation including temporal continuity, temporal periodicity, and spatial proximity are carefully chosen and simultaneously modeled by three different forms of graph Laplacian, named temporal graph Fourier transform, generalized temporal consistency regularization, and diffusion graph regularization. We then design an efficient solution algorithm via several effective numeric techniques to scale up the proposed model to network-wide kriging. By performing experiments on two public million-level traffic speed datasets, we finally draw the conclusion and find our proposed LETC achieves the state-of-the-art kriging performance even under low observation rates, while at the same time saving more than half computing time compared with baseline methods. Some insights into spatiotemporal traffic data modeling and kriging at the network level are provided as well.

Improvements in Resolution of ¹H NMR of solids.

Magic angle spinning (MAS) in 1H NMR has allowed progress from featureless spectra in static samples to linewidths of a few hundreds of Hertz for powdered solids at the fastest spinning rates available today (100-150 kHz). While this is a remarkable improvement, this level of resolution is still limiting to the widespread use of 1H NMR for complex systems. This review will discuss two recent alternative strategies that have significantly improved 1H resolution, when combined with fast MAS. The first is based on anti-z-COSY, a 2D experiment originally used for J decoupling in liquids, which removes residual broadening due to splittings caused by imperfect coherent averaging of MAS. The second strategy is to obtain pure isotropic proton (PIP) spectra in solids, by parametrically mapping any residual broadening due to imperfect averaging into a second dimension of a multidimensional correlation spectrum.

Machine learning for optical chemical multi-analyte imaging

Simultaneous sensing of metabolic analytes such as pH and O2 is critical in complex and heterogeneous biological environments where analytes often are interrelated. However, measuring all target analytes at the same time and position is often challenging. A major challenge preventing further progress occurs when sensor signals cannot be directly correlated to analyte concentrations due to additional effects, overshadowing and complicating the actual correlations. In fields related to optical sensing, machine learning has already shown its potential to overcome these challenges by solving nested and multidimensional correlations. Hence, we want to apply machine learning models to fluorescence-based optical chemical sensors to facilitate simultaneous imaging of multiple analytes in 2D. We present a proof-of-concept approach for simultaneous imaging of pH and dissolved O2 using an optical chemical sensor, a hyperspectral camera for image acquisition, and a multi-layered machine learning model based on a decision tree algorithm (XGBoost) for data analysis. Our model predicts dissolved O2 and pH with a mean absolute error of < 4.50·10−2 and < 1.96·10−1, respectively, and a root mean square error of < 2.12·10−1 and < 4.42·10−1, respectively. Besides the model-building process, we discuss the potentials of machine learning for optical chemical sensing, especially regarding multi-analyte imaging, and highlight risks of bias that can arise in machine learning-based data analysis.

Data-driven reliability assessment with scarce samples considering multidimensional dependence

This study proposes a data-driven method for assessing reliability, based on the scarce input dataset with multidimensional correlation. Since considering the distribution parameters estimated from the scarce dataset as those of the population may lead to epistemic uncertainty, the bootstrap resampling algorithm is adopted to infer the distribution parameters as interval parameters. To account for the variable dependence, vine copula theory is utilized to construct the joint probability density function (PDF) of input variables, and maximum likelihood estimation (MLE) and Akaike information criterion (AIC) analysis are employed to select optimal copulas based on the samples for the vine structure. Subsequently, the failure probability bounds of a response function are calculated based on the constructed joint PDF with interval distribution parameters by the active learning Kriging (AK) method combining the sparse grid integration (SGI) method. Finally, several examples are provided to demonstrate the feasibility and efficiency of the proposed method.

Asynchronising five-fold symmetry sequence for better homonuclear polarisation transfer in magic-angle-spinning solid-state NMR

Recoupling, decoupling, and multidimensional correlation experiments in magic-angle-spinning (MAS) solid-state NMR can be designed by exploiting the symmetry of internal spin interactions. One such scheme, namely, C521, and its supercycled version SPC521, notated as a five-fold symmetry sequence, is widely used for double-quantum dipole-dipole recoupling. Such schemes are generally rotor synchronised by design. We demonstrate an asynchronous implementation of the SPC521 sequence leading to higher double-quantum homonuclear polarisation transfer efficiency compared to the normal synchronous implementation. Rotor-synchronisation is broken in two different ways: lengthening the duration of one of the pulses, denoted as pulse-width variation (PWV), and mismatching the MAS frequency denoted as MAS variation (MASV). The application of this asynchronous sequence is shown on three different samples, namely, U–13C-alanine and 1,4-13C-labelled ammonium phthalate which include 13Cα-13Cβ, 13Cα-13Co, and 13Co–13Co spin systems, and adenosine 5′- triphosphate disodium salt trihydrate (ATP⋅3H2O). We show that the asynchronous version performs better for spin pairs with small dipole-dipole couplings and large chemical-shift anisotropies, for example, 13Co–13Co. Simulations and experiments are shown to corroborate the results.

Synergistic interplay of dual active sites on spinel ZnAl2O4 for syngas conversion

•Specific surface dual active sites on ZnAl2O4 for syngas conversion are identified •Dynamic interaction between the active sites and intermediates is elaborated •Synergistic interplay of heteroatoms for syngas conversion is revealed The urgent need for efficiency improvement in the oxide-zeolite bifunctional syngas-to-hydrocarbon catalysis necessitates in-depth mechanistic insights into this reaction, especially for the initial syngas conversion over the oxide component, which remains poor. Herein, we comprehensively investigated syngas conversion over a representative ZnAl2O4 spinel oxide with state-of-the-art solid-state NMR technologies. Notably, specific surface dual active sites for syngas activation with –AlIV–OH···ZnIII– structure were unambiguously identified. More importantly, the dynamic evolution of the reaction intermediates and active sites during the reaction process was elaborated at atomic level by a series of double resonance and multi-dimensional correlation NMR experiments. In combination with in situ spectroscopic characterizations, we revealed the full cycle of the formate-methoxy-based pathway for the syngas-to-methanol conversion via synergistic interplay of the dual active sites. The in-depth atomic-level understanding of the catalytic mechanism will be beneficial to further rational design of high-performance catalysts for syngas conversion. The urgent need for efficiency improvement in the oxide-zeolite bifunctional syngas-to-hydrocarbon catalysis necessitates in-depth mechanistic insights into this reaction, especially for the initial syngas conversion over the oxide component, which remains poor. Herein, we comprehensively investigated syngas conversion over a representative ZnAl2O4 spinel oxide with state-of-the-art solid-state NMR technologies. Notably, specific surface dual active sites for syngas activation with –AlIV–OH···ZnIII– structure were unambiguously identified. More importantly, the dynamic evolution of the reaction intermediates and active sites during the reaction process was elaborated at atomic level by a series of double resonance and multi-dimensional correlation NMR experiments. In combination with in situ spectroscopic characterizations, we revealed the full cycle of the formate-methoxy-based pathway for the syngas-to-methanol conversion via synergistic interplay of the dual active sites. The in-depth atomic-level understanding of the catalytic mechanism will be beneficial to further rational design of high-performance catalysts for syngas conversion.

Jointly Low-Rank Tensor Completion for Estimating Missing Spatiotemporal Values in Logistics Systems

With the deepening of industry 4.0 paradigm in logistics systems, artificial intelligent has been widely used to improve the quality of logistics services. Considering that data collected in comprehensive logistics service system usually integrate multistage complex information such as traffic flow records and spatiotemporal trajectory, it is inevitable that the data are incomplete and partially missing due to equipment failures, communication interruptions, etc. As an effective spatiotemporal completion tool in logistics systems, low-rank tensor completion has aroused extensive research interest thanks to its excellent performance on data recovery. Although existing tensor completion methods effectively capture the complex associations/dependencies of multidimensional inputs, they fail to exploit the potential characteristics of spatiotemporal data, such as the periodicity. In this article, we propose a jointly low-rank tensor completion method for logistics data completion, which constructs multiple periodic subtensors by setting an appropriate time window, then performs jointly low-rank completion and imputation. In addition, we also provide an optimization algorithm based on Alternating Direction Method of Multiplier framework for the proposed problem. Experimental results on four logistics-related datasets have further demonstrated the promising performance of the proposed method compared with other state-of-the-art competitors. We believe that the proposed approach not only effectively maintains the advantages of classical completion methods, but also fully excavates the multidimensional correlation and hidden patterns behind records, and further provides a novel and effective strategy for data completion and imputation in logistics systems.

Genome-wide analysis of epigenetic and transcriptional changes in the pathogenesis of RGSV in rice.

Rice grassy stunt virus (RGSV), a typical negative single-stranded RNA virus, invades rice and generates several disease signs, including dwarfing, tillering, and sterility. Previous research has revealed that RGSV-encoded proteins can force the host's ubiquitin-proteasome system to utilize them for viral pathogenesis. However, most of the studies were limited to a single omics level and lacked multidimensional data collection and correlation analysis on the mechanisms of RGSV-rice interactions. Here, we performed a comprehensive association analysis of genome-wide methylation sequencing, transcriptome sequencing, and histone H3K9me3 modification in RGSV-infested as well as non-infested rice leaves, and the levels of all three cytosine contexts (CG, CHG and CHH) were found to be slightly lower in RGSV-infected rice leaves than in normal rice. Large proportions of DMRs were distributed in the promoter and intergenic regions, and most DMRs were enriched in the CHH context, where the number of CHH hypo-DMRs was almost twice as high as that of hyper-DMRs. Among the genes with down-regulated expression and hypermethylation, we analyzed and identified 11 transcripts involved in fertility, plant height and tillering, and among the transcribed up-regulated and hypermethylated genes, we excavated 7 transcripts related to fertility, plant height and tillering. By analyzing the changes of histone H3K9me3 modification before and after virus infestation, we found that the distribution of H3K9me3 modification in the whole rice genome was prevalent, mainly concentrated in the gene promoter and gene body regions, which was distinctly different from the characteristics of animals. Combined with transcriptomic data, H3K9me3 mark was found to favor targeting highly expressed genes. After RGSV infection, H3K9me3 modifications in several regions of CTK and BR hormone signaling-related genes were altered, providing important targets for subsequent studies.

Sector rotation multi-factor stock selection strategy based on six boards

This paper proposes an investment strategy based on a regression linear model to predict sector rotation and a multi-factor scoring model to select stocks to achieve excess returns. The strategy uses January 2006 December 2020 as the model construction period and January 2021 June 2021 as the model validation and backtesting period. Using Shenyin &amp; wanguo’s industry classification as a standard, the cyclicality of industries is classified by β value and the existence of sector rotation is verified using the difference in rate of return between cyclical and non-cyclical industry. The factors associated with the annualized rate of return are selected and the fitted formula for the annualized rate of return is obtained by regressing linear equation and neural network multidimensional correlation to select the dominant board. Valid factors for each board were selected for fitting, and the factors were assigned weights for scoring, resulting in 32 stocks. The backtest results significantly outperformed the CSI 300 Index, and the model performance is stable and can achieve better returns.

Impact of the COVID-19 pandemic on the socio-economic development of the region

The article presents an analysis of the consequences of economic instability caused by the COVID‑19 pandemic for the socio‑economic development of the region. The study was carried on the example of the Sverdlovsk region according to the Federal State Statistics Service. The article sets and solves three research problems. First, to assess how the overall socio‑economic condition of the region has changed because of the COVID‑19 pandemic. Second, to identify the impact of the COVID‑19 pandemic on the nature and structure of the relationship between the socio‑economic indicators of the region. Third, to analyze the impact of the COVID‑19 pandemic on the trends in the socio‑economic indicators of the region, considering their possible delay and seasonal fluctuations. The first task is solved using statistical methods of pattern recognition, the second task — by methods of multivariate correlation analysis, and for the third task stochastic time series models are used. The largest contribution to the division into clusters was made by the number of births, the turnover of public catering and the volume of goods, works and services performed on their own; their dynamics most clearly traces the impact of the COVID‑19 pandemic in the Sverdlovsk region. Multidimensional correlation analysis established that because of the introduction of restrictive and supportive measures, multidirectional trends in the socio‑economic indicators of the region were formed, which violated the previously established relationships between them. Based on the obtained coefficients of the autoregressive model, one can find a certain pattern consisting in the growth of the influence of short‑term (lag 1 and 3 months) and the decrease of the long‑term influence (lag 12 months) change in indicators. According to the calculations obtained, the economy of the Sverdlovsk region during the analyzed period failed to overcome the destabilizing effect of the consequences of the COVID‑19 pandemic.

A Novel Nonconvex Low-Rank Tensor Completion Approach for Traffic Sensor Data Recovery From Incomplete Measurements

Complete traffic sensor data is considered to be one of the critical ingredients for intelligent transportation systems (ITS). However, the traffic measurements prevalently suffer from the inevitable and ubiquitous missing values. Current missing data completion algorithms are difficult to leverage the global low-rank property and the fine-grained spatial-temporal structure simultaneously. To circumvent this problem, this work presents a novel collaborative nonconvex low-rank spatial-temporal data tensor completion model that can take full advantage of the inherent spatial-temporal characteristics of traffic measurement data. First, the tensor Schatten p-norm, as an effective nonconvex surrogate of rank function, is used to exploit the global multi-dimensional correlation of traffic data. Then, we present an elastic net self-representation method and utilize an autoregression model in order to simultaneously capture the self-similarity and the temporal continuity of traffic data acquired in the same sensor network. By integrating the above elements in a unified nonconvex learning model, our method can explore the inherent structure of traffic data from the viewpoints of both global multi-dimensional correlation and fine-grained spatial and temporal dependency. Then, in the view of the general framework of the alternating directions method of multipliers (ADMM), an efficient iterative algorithm is designed to solve our model. Besides, to optimize the parameter combination of the model, a Bayesian optimization-based parameter selection algorithm is developed, which avoids manual parameter adjustment. Extensive experiments and analyses on two real-world traffic datasets are carried out. The results demonstrate the advantages of our model under diverse missing patterns and missing ratios.

Energy cost optimization of globally distributed Internet Data Centers by copula-based multidimensional correlation modeling

The high operating costs of Internet Data Centers (IDC) are a major challenge for their owners worldwide. Therefore, more attention has recently been paid to the energy and cost management of IDCs. This paper investigates the optimal operational strategy for minimizing the electricity costs of a group of globally distributed IDCs in different locations under various day-ahead electricity markets, and each is equipped with a high-performance energy storage system. For this goal, optimal workload dispatching and optimal energy management of the storage units of all IDCs are simultaneously perused by the proposed problem. The system is modeled regarding power balancing constraints, battery costs, and quality of service (QoS). For more practical results, a penalty function is also considered when QoS constraints are not perfectly met, and the impact of the batteries’ depth of discharge on the cost of energy storage is also modeled. Moreover, the cross-correlations between the traffic of IDCs are also considered by the multidimensional copula function. The proposed energy cost optimization is linearized for increasing the accuracy of convergence. The results show that not only the power consumption pattern of the IDCs is significantly improved, but also the cost of power consumption is reduced by 34%. The results also prove the positive effect of battery discharge on workload dispatch and represent a compromise between battery costs and electricity cost savings.

"Multidimensional correlation analysis of temperature and contact time on eradication of biofilms of Cronobacter sakazakii on abiotic surfaces by combination of hypochlorite and malic acid".

In the present study, malic acid in combination with sodium hypochlorite is evaluated for eradication of biofilms formed by Cronobacter sakazakii strains individually and in a cocktail on different abiotic surfaces. The biofilm formation by five strains of C. sakazakii and their cocktail culture on different substrates was studied in Tryptone Soy Broth (TSB) and reconstituted Powdered Infant Formula (PIF). Further, the effect of temperature (4, 27, 37 and 50°C) and contact time (10, 20, 30, 40, 50 and 60min) on antibiofilm potential of test solution (0.0625mol l-1 malic acid and 0.00004mol l-1 sodium hypochlorite) against biofilm formed by C. sakazakii cocktail culture was investigated on these surfaces. The effect was evaluated in terms of viable cell count and biofilm texture using scanning electron microscopy (SEM). Principal Component Analysis (PCA) revealed that the maximum biofilm reduction was observed for stainless steel at 4°C after 60min of contact whereas at 25, 37 and 50°C, maximum biofilm reduction was observed for polycarbonate. For glass and polyurethane, maximum log reductions were observed at 50°C. The SEM images revealed cell surface deformation and disruption in biofilms after treatment with the test solution. The antibiofilm potential was observed to be greatly affected by contact time and temperature. These results indicated that the combination of malic acid NaOCl can effectively kill and remove C. sakazakii biofilms from food contact surfaces and enteral feeding tubes.

Built environment and travel behavior in rural areas: A scientometric literature review

With the rise of global urbanization, the rural built environment has undergone tremendous changes. As such, the rural built environment impacts on residents’ daily travel behavior is getting more researchers’ attention. To date, most of the research focuses on urban areas in developed countries. To understand the state-of-the-art of interplay between the rural built environment and travel behaviors and to identify future research directions, this study adopts a science mapping approach to identify the relevant topics, authors, journals, and countries of the research done. This study proceeds through bibliometric retrieval of articles from 2005, followed by scientometric analysis and qualitative discussion. 37 documents are found to compare urban and rural domains, with 28 on the rural built environment. Research gaps and the research trends are discussed, of which the main themes are multi-dimensional correlation comparison of rural transportation service systems and emerging transportation modes, the influence of rural social and cultural factors on travel behavior, and low-carbon sustainable transportation. This review provides empirical foundation for current state-of-the-art and identifies the future research directions, specifically for rural built environment impact on travel behavior.

Projections of two-particle correlations onto transverse rapidity in Au+Au collisions at sNN=200 GeV at STAR

Two-particle correlation measurements projected onto two-dimensional, transverse rapidity coordinates (yT1,yT2) provide an independent, orthogonal view of the multidimensional correlation distribution that is most often studied via angular projections. As such, these independent transverse projections allow access to manifestations of dynamical fluctuations in relativistic heavy-ion collisions that angular-correlation measurements may not be sensitive to. We report nonidentified charged-particle correlations for Au+Au minimum-bias collisions at sNN=200 GeV taken by the STAR experiment at the Relativistic Heavy-Ion Collider (RHIC). Correlations are presented as two-dimensional functions of transverse rapidity for like-sign, unlike-sign, and all charged-particle pairs, as well as for particle pairs whose relative azimuthal angles lie on the near-side, the away-side, or at all relative azimuth. The correlations are constructed using charged particles with transverse momentum pT≥0.15 GeV/c, pseudorapidity from −1 to 1, and azimuthal angles from −π to π. The significant correlation structures that are observed evolve smoothly with collision centrality. The major correlation features include a saddle shape plus a broad peak with maximum near yT≈3, corresponding to pT≈1.5 GeV/c. The broad peak is observed in both like- and unlike-sign charge combinations and in near- and away-side relative azimuthal angles. The all-charge, all-azimuth correlation measurements are compared with the predictions of hijing and epos to provide theoretical context for these new measurements. The results indicate that the correlations for peripheral to mid-central collisions can be approximately described as a superposition of nucleon + nucleon collisions with minimal effects from the quantum chromodynamics medium. Strong medium effects are indicated in mid- to most-central collisions.1 MoreReceived 22 April 2022Accepted 28 September 2022DOI:https://doi.org/10.1103/PhysRevC.106.044906©2022 American Physical SocietyPhysics Subject Headings (PhySH)Research AreasParticle correlations & fluctuationsRelativistic heavy-ion collisionsNuclear Physics

A Novel Spatiotemporal Data Low-Rank Imputation Approach for Traffic Sensor Network

The Internet of Things (IoT) has enormous potential to transform the transport industry by improving passenger experiences, safety, and efficiency. However, the collected spatiotemporal data by traffic sensor network often suffer from missing values (MVs), which affect the overall performance of the system. As a result, accurate recovery of MVs is essential for the successful application of IoT in transportation. In this article, we propose a novel MVs imputation model by integrating low-rank tensor completion (LRTC) and sparse self-representation into a unified framework. In doing so, the global multidimensional correlation, as well as sample self-similarity, can be well leveraged for imputation. In order to solve the proposed model, an elaborate solution algorithm is developed, following the principle of alternating direction method of multipliers (ADMMs). Importantly, each step in ADMM can be implemented efficiently by analyzing the problem structure. Moreover, in order to select proper parameters for the model, an improved harmony search heuristic algorithm based on dual harmonies generation strategy is developed, thus sufficiently considering the information contained in current harmony memory. The experiments on two real-world traffic data are carried out to evaluate the proposed approach. The results verify that in comparison with the classic matrix/tensor completion and other competing algorithms, our method significantly improves the imputation performance.

Photon Correlations in Spectroscopy and Microscopy

Measurements of photon temporal correlations have been the mainstay of experiments in quantum optics. Over the past several decades, advancements in detector technologies have supported further extending photon correlation techniques to give rise to novel spectroscopy and imaging methods. This Perspective reviews the evolution of these techniques from temporal autocorrelations through multidimensional photon correlations to photon correlation imaging. State-of-the-art single-photon detector technologies are discussed, highlighting the main challenges and the unique current perspective of photon correlations to usher in a new generation of spectroscopy and imaging modalities.

An advanced mixed-degree cubature formula for reliability analysis

Efficient assessment of mechanical system reliability subject to arbitrary probability distributions and dependent input parameters signifies an important yet challenging task. To tackle this problem, this study proposes a new moment-based method for reliability analysis of complex mechanical systems which incorporates the advanced mixed-degree cubature formula and vine copula function. To start with, a method integrating the vine copula function and Rosenblatt transformation is developed for transferring the uncertainty with dependent random variables, in which the vine copula is applied for depicting the correlation between random variables. An advanced mixed-degree cubature formula is then established for calculating statistical moments of performance function, which enables to capture sufficient uncertainty information of variables after rotating integral node. The Hermite polynomial model is adopted here to reconstruct the probability distribution of performance function with the statistical moments. To demonstrate the effectiveness of the proposed method, four illustrative examples are implemented in this study, in which Monte Carlo Simulation (MCS) and dimension-reduction techniques, are performed for comparisons. The results show that the proposed method can handle multi-dimensional correlation effectively and well balance computational accuracy and efficiency for both statistical moments and probability distribution evaluations.

Identification and Characterization of Genes Related to the Prognosis of Hepatocellular Carcinoma Based on Single-Cell Sequencing.

The heterogeneity of hepatocellular carcinoma (HCC) highlights the importance of precision therapy. In recent years, single-cell RNA sequencing has been used to reveal the expression of genes at the single-cell level and comprehensively study cell heterogeneity. This study combined big data analytics and single-cell data mining to study the influence of genes on HCC prognosis. The cells and genes closely related to the HCC were screened through single-cell RNA sequencing (71,915 cells, including 34,414 tumor cells) and big data analysis. Comprehensive bioinformatics analysis of the key genes of HCC was conducted for molecular classification and multi-dimensional correlation analyses, and a prognostic model for HCC was established. Finally, the correlation between the prognostic model and clinicopathological features was analyzed. 16,880 specific cells, screened from the single-cell expression profile matrix, were divided into 20 sub-clusters. Cell typing revealed that 97% of these cells corresponded to HCC cell lines, demonstrating the high specificity of cells derived from single-cell sequencing. 2,038 genes with high variability were obtained. The 371 HCC samples were divided into two molecular clusters. Cluster 1 (C1) was associated with tumorigenesis, high immune score, immunotherapy targets (PD-L1 and CYLA-4), high pathological stage, and poor prognosis. Cluster 2 (C2) was related to metabolic and immune function, low immune score, low pathological stage, and good prognosis. Seven differentially expressed genes (CYP3A4, NR1I2, CYP2C9, TTR, APOC3, CYP1A2, and AFP) identified between the two molecular clusters were used to construct a prognostic model. We further validated the correlation between the seven key genes and clinical features, and the established prognostic model could effectively predict HCC prognosis. Our study identified seven key genes related to HCC that were used to construct a prognostic model through single-cell sequencing and big data analytics. This study provides new insights for further research on clinical targets of HCC and new biomarkers for clinical application.

Adaptive Band Extraction Based on Low Rank Approximated Nonnegative Tucker Decomposition for Anti-Friction Bearing Faults Diagnosis Using Measured Vibration Data

Condition monitoring and fault diagnosis are topics of growing interest for improving the reliability of modern industrial systems. As critical structural components, anti-friction bearings often operate under harsh conditions and are contributing factors of system failures. Efforts have been cast on bearing diagnostics under the sensor fusion and machine learning framework, whilst challenges remain open on the identification of incipient faults. In this paper, exploiting multi-way representations and decompositions of measured vibration data, a novel band separation method based on the factorization of spectrogram tensors using the low rank approximated nonnegative Tucker decomposition (LRANTD) is proposed and applied to identify detailed fault signatures from the spectral, temporal, and spatial dimensions, flexible for extracting multi-sensor features and multi-dimensional correlations. With the proposed method, informative frequency bands of the latent vibrational components can be automatically extracted, in accordance with the inherent temporal patterns that can be conveniently fed for spectral analysis and fault discrimination. Furthermore, an improved cross-spectrum can be calculated from multi-channel vibrations via LRANTD with enhanced fault features. Based on the real-world vibration data of the accelerated bearing life tests, detailed experimental studies and thorough comparisons to the conventional benchmarks have verified the effectiveness of the reported diagnostic methodology. The proposed method significantly improves the presence of the bearing frequency peaks distinctly over the background noises in the spectrum and hence improves the bearing defect detection process.